And then they were eight. Tracking polls that is. Which one should you follow? Or, rather, which should you discount? Happily, Nate Silver is here to explain it all.
What I find odd about American polls, mind you, is their tiny sample size. National Journal's tracking poll only follows a few hundred people for instance and even national opinion polls often only consult around 1,000 people. That's roughly the same kind of sample size used by British pollsters. But of course there are 60 million people in Britain and 300 million in the United States. Our polls should, therefore, be more accurate. Right? And in America how difficult would it be to poll 5,000 people instead of just a few hundred?
UPDATE: As commenters point out, I don't know very much about polling. Or statistics.